Gilligan’s Ethics of Care: S-Tier Behavioral Designer’s Guide

In the late 1970s, two eleven-year-olds named Jake and Amy were asked the same question: should a man named Heinz steal an overpriced drug to save his dying wife? Jake treated it, in his own words, sort of like a math problem with humans. Life beats property, so Heinz steals, and a sensible judge will go easy on him. Amy refused the equation. If Heinz steals, he might go to prison, and then who takes care of his wife when she gets sicker? Shouldn’t Heinz and the druggist just talk it out? On Lawrence Kohlberg’s famous scale of moral development, Jake scored a full stage above Amy.

Carol Gilligan looked at those transcripts and asked the question that would split moral psychology in two: what if Amy isn’t behind? What if she’s answering a different question, in a different voice, that the scoring manual was never built to hear? Jake heard “which rule wins?” Amy heard “how do these people keep caring for each other through this?” One of those questions had a fifty-year-old measurement tradition behind it. The other had been filed under “immature.”

I have spent more than twenty years studying what moves people to act, and I built the Octalysis Framework to map the eight Core Drives underneath that action. Here is why Gilligan matters to me as a behavioral designer, and why she should matter to you: almost every engagement dashboard ever built scores like Kohlberg. Points, ranks, streaks, achievements, fairness, individual progress. The justice voice. Meanwhile the force that quietly decides whether people stay, in products, in communities, in jobs, in marriages, is usually the other voice: who depends on me here, who do I care for, who would notice if I left. This post walks through what Gilligan actually claimed, where the evidence pushed back, what empathy physically looks like in the brain, and then the part almost nobody designs on purpose: how to build experiences where care is the engine instead of the afterthought.

Speed Run Notes

• Gilligan’s ethics of care says moral life runs on two voices: justice (rights, rules, autonomy) and care (relationships, responsibility, response). Psychology scored only the first and called the second immature.
• Her three stages of care: survival (care for self), self-sacrifice (care for others at your expense), and mature care (self and others in one web). The two transitions between them are where the real moral work happens.
• The empirical record is humbling: meta-analyses find only small, inconsistent gender differences in moral orientation. The care voice is human, not female. That correction made the theory more useful, not less.
• Most “social” features in products are justice-frame: leaderboards, status, fairness. Care-frame design, where someone depends on you and you respond, is rarer and retains far harder.
• The same care mechanic splits White Hat from Black Hat: reflecting real mutual care builds loyalty that compounds; manufacturing guilt and dependency (the sad owl) builds a retention curve with resentment baked in.
• The S-Tier move: stop designing only for the lone achiever. Design who your user gets to care for, then close the loop so the care is received and answered, not just performed into a void.

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

TEDxLausanne · Gamification to Improve Our World17 min

What Is the Ethics of Care?

The ethics of care is a moral framework, introduced by psychologist Carol Gilligan in her 1982 book In a Different Voice, which holds that moral problems are best understood as problems of relationship and responsibility rather than problems of competing rights and abstract rules. Where traditional moral theories ask “what is fair?” and “whose rights win?”, the care perspective asks “who is affected?”, “what do they need?”, and “how do I respond without abandoning anyone, including myself?”

Gilligan developed the theory as a direct challenge to her mentor. Lawrence Kohlberg’s stage model of moral development, built from a longitudinal study that began with 84 boys and no girls, defined moral maturity as the climb toward abstract, universal principles of justice. When women were scored on Kohlberg’s interviews, they landed disproportionately at stage three, the “good interpersonal relationships” stage, and the field’s quiet conclusion was that women tended to be morally less developed. Kohlberg and Kramer’s 1969 paper said this almost out loud.

Gilligan’s counterclaim was surgical. The instrument was not measuring deficient moral reasoning. It was measuring a different moral logic with a ruler built for the first one. The justice voice reasons from separation: individuals with rights, conflicts settled by ranking principles. The care voice reasons from connection: selves embedded in relationships, conflicts settled by tracing who needs what and who can respond. Neither is a degraded copy of the other. A moral agent who can only do one of them is incomplete.

Two details about the theory get lost in summaries, and both matter for anyone who designs for humans. First, Gilligan never claimed care is women’s morality and justice is men’s. She claimed the care voice was culturally associated with women and therefore systematically unheard, which is a claim about whose answers get scored, not about chromosomes. Second, the ethics of care is developmental. It is not a warm feeling. It has stages, transitions, and failure modes, and the failure modes (especially self-erasing sacrifice) do real damage. That structure is what makes it usable for design rather than decoration.

Carry one framing into the rest of this guide: justice and care are not demographic labels, they are two questions a mind can ask about the same situation. “What is fair here?” and “Who needs what here?” Every user you design for asks both, in proportions that shift with context, stakes, and how the situation is framed. A design, like a scoring manual, chooses which question it can hear. Most products built in the last twenty years can only hear the first one, and the rest of this pillar is about what becomes possible when you build for the second.

Gilligan’s Three Stages of Care

Gilligan first laid out the sequence in her 1977 Harvard Educational Review article, drawing on interviews with women making real decisions under real stakes. The progression is not about caring more; it expands who counts as a legitimate object of care, and the two transitions between stages carry most of the weight.

Stage 1: Orientation to Individual Survival

The first stage is care aimed entirely at the self. The moral question is “what do I need to be okay?”, and other people enter the picture as resources or threats. Gilligan did not treat this as villainy. The women in her studies who reasoned this way were usually the ones with the fewest options and the least support, and self-protection was a rational response to a world that had not been safe. Survival reasoning is what care looks like when the web of relationships has failed you.

The first transition, from selfishness to responsibility, happens when a person starts to see that they are attached to others and that their choices land on those others. The vocabulary shifts from “what I want” to “what I should do.” This is the entry into conventional moral life, and most people make it.

Stage 2: Goodness as Self-Sacrifice

The second stage equates being good with taking care of everyone else, at whatever cost to yourself. Your own needs get reclassified as selfish by default. Conflict gets resolved by you absorbing the damage. Gilligan heard this voice constantly in her interviews: women who could argue eloquently for anyone’s needs except their own, and who experienced saying “I want” as a moral violation.

This is the stage conventional femininity rewarded, and Gilligan’s treatment of it is one of the most underrated pieces of the theory. She refused to romanticize it. Stage two care is unstable and quietly dishonest, because the carer’s real needs do not vanish; they leak out as resentment, martyrdom, and manipulation. A morality that requires your own erasure is not goodness. It is a developmental waystation that a lot of people, and a lot of institutions, get stuck in.

The second transition, from goodness to truth, is the hard one. It happens when a person admits that excluding themselves from their own circle of care is itself a moral failure, and that “is this honest?” matters as much as “is this nice?” Gilligan watched women in her abortion-decision study make exactly this move under pressure: stop asking “what would a good woman do?” and start asking “what is actually true about my situation and everyone in it, me included?”

Stage 3: The Mature Ethics of Care

The third stage holds self and others in a single web. The governing principle becomes nonviolence: hurt no one, where “no one” finally includes you. Care at this stage is no longer compliance or performance. It is a chosen responsibility, exercised with honesty about real needs and real limits. The mature carer can say no, can disappoint someone for the sake of a deeper truth, and can receive care without reading it as weakness.

Notice the shape of the whole arc: care for self, then care for others, then care for the relationship between them. It is a dialectic, not a ladder of niceness. And here is the design takeaway I will build on later in this guide: most products that attempt “social good” mechanics are accidentally built for stage two. They reward visible sacrifice and unpaid emotional labor, then act surprised when their most generous users burn out and leave. A care system designed for stage three looks different. It builds the carer’s replenishment into the loop.

The Different Voice: How Gilligan Built the Theory

Gilligan’s evidence base was three interview studies, run in the 1970s while she was teaching alongside Kohlberg at Harvard. The college student study followed students who had taken a course on moral and political choice, and tracked how they reasoned about their own lives five years later. The abortion decision study interviewed 29 women who were actually deciding whether to continue a pregnancy, which gave Gilligan something Kohlberg’s hypothetical dilemmas never had: moral reasoning with skin in the game. The rights and responsibilities study sampled males and females across nine age points, matched for age, intelligence, education, and social class, and asked them to reason about both hypothetical and real dilemmas.

Out of those transcripts came the book’s central exhibits. Jake and Amy on Heinz is the famous one, and it is worth restating what Gilligan actually argued, because the popular version flattens it. Her point was never that Amy’s answer was superior. Her point was that Kohlberg’s scoring manual could only see Amy’s answer as a failed attempt at Jake’s. Amy’s response (“if he steals the drug, he might have to go to jail, and then his wife might get sicker again, and he couldn’t get more of the drug”) is not weak logic. Amy is tracking consequences across a web of relationships over time, and refusing the dilemma’s premise that someone must lose. The manual heard evasion. Gilligan heard a different question being answered.

The abortion study gave the theory its developmental spine. Listening to women reason through an irreversible decision that pitted their own needs against others’ expectations, Gilligan heard the three-stage progression: survival reasoning, self-sacrificial goodness, and a hard-won honesty that included the self in the moral equation. The word “selfish” appeared in those interviews with remarkable frequency, almost always as the charge the women were trying to escape. The entire second transition is visible in how that word loses its grip.

One more piece of the method deserves attention from anyone who works with user research: Gilligan’s core instrument was listening for the question behind the answer. Two people can produce the same surface choice (steal the drug, don’t steal it) from entirely different moral architectures. If your interview rubric only codes the choice, you will never see the architecture. That insight transfers directly to product research, where “users who churned” is a single bucket hiding people who left because they finished, people who left because they were overwhelmed, and people who left because nobody there needed them.

The Listening Guide: A Method Worth Stealing

Gilligan’s later methodological work with Lyn Mikel Brown, published in their 1992 book Meeting at the Crossroads, turned that instinct into a formal technique called the Listening Guide. The same transcript gets read multiple times, each pass tuned to a different frequency: once for the plot, once for the speaker’s first-person “I” statements pulled out and read as a poem in sequence, and then once for each contrapuntal voice, the care voice and the justice voice, marked in different colors wherever they surface. The premise is that people are polyphonic. A single answer can carry a confident justice melody with a care countermelody underneath it, and a rubric that codes only the loudest line will report a person who does not exist.

I want every researcher who runs user interviews to sit with that design. Our standard coding passes (tag the feature request, tag the complaint, tag the jobs-to-be-done) are single-voice instruments. The churned user who says “it just wasn’t worth the money” in the plot layer is often saying “nobody there knew me” two layers down, in the I-statements. You do not need Gilligan’s politics to use Gilligan’s method: read your transcripts more than once, listen for the voice the rubric was not built to hear, and assume the most decision-relevant signal is the one your current instrument codes as noise.

What Gilligan Got Right

Four of Gilligan’s claims have aged well, and they are the reasons this pillar exists.

The measurement critique was correct. Building a universal theory of moral development from an all-male longitudinal sample, then scoring women as deficient against it, is a textbook sampling failure. Whatever you conclude about gender and care, the methodological point stands on its own and applies far beyond moral psychology. Every “universal” engagement benchmark built from one user population makes the same mistake.

The two orientations are real. Subsequent research confirmed that people really do frame moral problems in terms of justice and rights, or in terms of care and responsibility, and that the framing changes what they notice, what they count as a solution, and what they regret afterward. Lyons’ coding work and decades of dilemma studies support the distinction itself, whoever uses it. The two voices themselves survived every fight about who speaks them.

She caught the pathology of stage two. Gilligan’s refusal to equate care with self-sacrifice was decades ahead of its time. Caregiver burnout research, codependency literature, and every community manager who has watched their most devoted volunteers flame out are all rediscovering her second transition: care that excludes the carer is not sustainable and not even honest.

She named what institutions can’t see. Justice reasoning produces artifacts that institutions can process: rules, scores, precedents, rankings. Care reasoning produces outcomes that mostly look like nothing happening: the conflict that never escalated, the member who didn’t quit, the wife who was not left alone. Gilligan saw that what a measurement system cannot represent, it will eventually call worthless. Forty years later, every analytics dashboard in the world is still proving her right.

Where the Ethics of Care Falls Apart

I teach frameworks for a living, and the fastest way to lose a sophisticated audience is to present one without its bruises. The ethics of care has three big ones.

The Gender Difference Mostly Didn’t Replicate

The strong empirical reading of In a Different Voice (women reason in care, men in justice) collapsed under meta-analysis. Walker’s 1984 review found that sex differences on Kohlberg’s own measures were rare and usually vanished once education and occupation were controlled. Thoma’s 1986 meta-analysis found differences on the Defining Issues Test that were trivially small, and where they existed, they slightly favored women. The decisive blow came from Jaffee and Hyde’s 2000 meta-analysis of 113 studies: a small effect for care orientation favoring women (d = -0.28) and a small effect for justice orientation favoring men (d = 0.19), with most of the variance living inside each gender, not between them. Both sexes use both orientations, and the dilemma’s content predicts the voice better than the speaker’s gender does. If you want the defensible version of Gilligan, it is this: the care voice is real, undervalued, and human. The “different voice” belongs to a perspective, not a sex.

The Essentialism Trap

Feminist scholars landed the second punch, and it is philosophically the most interesting one. If you celebrate care as women’s distinctive moral voice, you have handed a polished new justification to every structure that assigns women the unpaid caring work. Catharine MacKinnon made the sharpest version of the argument: the “different voice” may be the voice of subordination itself, the moral style of people who historically could not afford to claim rights and had to manage relationships to survive. Praising it without asking where it came from risks consecrating the conditions that produced it. Gilligan’s defenders reply that she described association, not essence, and her own later work pushed the same direction. But the trap is real, and any designer who builds “community care” features that quietly route all the emotional labor to the same demographic of users is walking straight into it.

A Standpoint, Not (Yet) a Theory

The third critique is structural. Justice ethics arrives with decision procedures: rank the principles, apply the rule, defend the verdict. Early care ethics arrived with an orientation and a warning, which made it hard to operationalize and easy to dismiss as vibes. When care obligations conflict (my child versus my patient, my team versus my own limits), Gilligan’s psychology describes the struggle but does not adjudicate it. It took philosophers to build the load-bearing structure: Nel Noddings’ analysis of the caring relation as engrossment plus motivational displacement, completed only when the cared-for receives it, and Joan Tronto’s four phases of care (caring about, taking care of, care-giving, and care-receiving, each with its own virtue and its own failure mode). Tronto’s fourth phase will do the most work later in this guide: care is not complete until the response comes back. Care without a receipt loop is just output.

What’s Really Happening Inside the Brain

Gilligan built the theory from interview transcripts. Neuroscience later went looking for the two voices, and what it found is more interesting than a simple confirmation.

Joshua Greene’s fMRI work at Princeton put people in a scanner with trolley-style dilemmas and found a dual-process architecture: impersonal dilemmas (pull a lever) preferentially recruited dorsolateral prefrontal regions associated with controlled, calculation-style reasoning, while personal dilemmas (push the man yourself) lit up medial prefrontal cortex, posterior cingulate, and emotion-processing regions. That is not literally justice versus care, but it shows the brain runs at least two distinguishable moral pipelines, one cold and rule-friendly, one hot and relationship-sensitive, and which one fires depends on how the problem is framed. Gilligan’s claim that the question’s framing selects the voice has a physical footprint.

The care side has its own circuitry. Tania Singer’s empathy studies showed that watching a loved one in pain activates the anterior insula and anterior cingulate cortex, the same affective regions that fire for your own pain. Your brain runs a low-resolution copy of the suffering of people you are attached to, which is the biological floor under Amy’s reasoning: the wife’s sickness and Heinz’s prison sentence are not abstractions to a care-oriented mind; they are simulated in tissue. Decety and Jackson’s work mapped the full architecture: shared affect, self-other distinction, and regulation, working as separable components.

The most design-relevant finding is the empathy-compassion split. Klimecki and Singer trained people in empathic resonance versus compassion (warm, action-oriented concern), and the two produced different signatures: empathy training alone increased anterior insula activity and reported distress, while compassion training engaged medial orbitofrontal and striatal reward circuitry and increased positive affect and helping. Pure empathic resonance burns the carer; compassion sustains them. There, in the scanner, is Gilligan’s second transition. Stage two (absorb everyone’s pain) is neurologically a distress state. Stage three (respond from a stable self) runs on reward circuitry and lasts. The mature ethics of care is not a moral nicety but the only configuration of caring the nervous system can run indefinitely.

The Ethics of Care vs Other Theories

Ethics of Care vs Kohlberg’s Moral Development

This is the founding rivalry, and the cleanest way to hold it: Kohlberg’s stage model maps the development of justice reasoning, from punishment-avoidance through social contracts to universal principles. Gilligan maps the development of care reasoning, from survival through self-sacrifice to mature interdependence. Kohlberg measures how people rank principles when interests collide. Gilligan measures how people sustain relationships when needs collide. The modern reading is that the two are parallel tracks rather than competitors, and a complete moral agent (or a complete product team) needs both: justice to keep the care from collapsing into favoritism, care to keep the justice from collapsing into bookkeeping.

Ethics of Care vs Haidt’s Moral Foundations

Jonathan Haidt’s Moral Foundations Theory actually enshrines Gilligan’s territory as its first foundation: Care/Harm, sitting alongside Fairness, Loyalty, Authority, and Sanctity. The deep difference is architectural. Haidt treats care as one of several evolved intuitions that fire automatically and get rationalized afterward. Gilligan treats care as a developmental achievement with stages, meaning it can mature, stall, or deform. Haidt tells you care exists in everyone’s moral taste profile. Gilligan tells you what skilled care looks like versus the self-erasing counterfeit, which is exactly the distinction a designer needs and a taste profile cannot give.

Gilligan vs Noddings: The Psychologist and the Philosopher

Nel Noddings’ 1984 book Caring took the orientation Gilligan had documented and built the philosophy. Her central move was relocating the unit of analysis: the moral atom is not the carer’s virtue but the caring relation, which exists only when the one-caring attends and the cared-for receives and responds. Unreceived care is incomplete care, however noble it feels from the inside. For anyone building products, that is the single most actionable sentence in the entire literature, and I will cash it out in the Octalysis section below: most “caring” features ship the giving half of the loop and forget the receiving half.

Ethics of Care vs Virtue Ethics

Care ethics is often filed under virtue ethics, since both reject rule-first morality in favor of character and context. The fit is imperfect in one telling way. Aristotle’s virtues live inside the flourishing individual; care ethics’ central good lives between individuals, in the relation. Virtue ethics asks “what kind of person should I be?” Care ethics asks “what does this relationship need from me now?” A leaderboard can, at a stretch, gamify virtue. It takes a different design entirely to gamify a relation, which is precisely why so few products have done it.

The Ethics of Care in the Real World

Healthcare: The Field That Institutionalized Care

Nursing is the one profession that took care ethics and wrote it into its codebase. Patient-centered care, family-witnessed resuscitation, hospice philosophy: these are stage-three structures, built on the premise that the unit of treatment is a person embedded in relationships, not a diagnosis with a room number. The instructive failure lives next door: systems that score hospitals on throughput and procedure counts while the relational work (the nurse who notices a patient stopped eating, the consultant who phones a frightened family) is invisible in the metrics and therefore perpetually understaffed. That split, the justice voice running procurement while the care voice runs the night shift, is Gilligan’s two-voice diagnosis at industrial scale.

The Workplace: Glue Work and the Invisible Economy

Every organization runs on a layer of uncounted care: the engineer who keeps the onboarding doc alive, the manager who notices someone drowning before they say it, the colleague who de-escalates the meeting nobody else could sit through. Engineering leader Tanya Reilly gave it a name, glue work, and pointed out the trap Gilligan would recognize instantly: the work is essential, disproportionately routed to women, and absent from promotion rubrics, which only parse individual, legible achievement. The justice-voice instrument (the performance review) cannot see the care-voice contribution, so organizations systematically promote the Jakes and burn out the Amys. The fix is not pity but measurement reform: make the relational labor legible, then reward it.

Education: Noddings’ Classroom

Noddings spent her career arguing that students learn from people they are in caring relations with, and that the relation, not the content delivery, is the channel. The research on teacher-student relationships keeps backing her: perceived teacher care predicts engagement and persistence, especially for students with the least support elsewhere. Notice the Tronto loop in any good classroom: the teacher attends (caring about), takes responsibility (taking care of), teaches responsively (care-giving), and then, the step weak teachers skip, watches whether it landed and lets the student’s response reshape the next move (care-receiving). Adaptive learning software that personalizes difficulty is automating phases one through three. Almost none of it closes phase four.

Products and Communities: Where Care Already Retains

Look at the products people refuse to abandon and you keep finding care loops where the engagement charts only show “social features.” Animal Crossing players log in because villagers miss them. Guild healers in World of Warcraft stay subscribed years past their own boredom because the raid needs them on Tuesday. Wikipedia’s editors do unpaid maintenance work because the commons depends on someone. CaringBridge hosts some of the most loyal usage on the internet around literally nothing but a care loop. And Finch, the self-care pet that grows when you look after yourself, is the rare mainstream app built on Gilligan’s third stage: it points the care loop back at the carer without letting it collapse into either selfishness or sacrifice. None of these are leaderboard psychology. All of them are somebody-depends-on-me psychology, and it is the stickiest force in product design that almost no metrics suite can see.

The Elephant in the Room

Here is what the engagement industry will not put on a slide: nearly every metric we optimize is a justice-voice metric, and we chose it because it is easy to count, not because it is what keeps people.

Daily actives, streaks, points, conversion, individual progression: all of it models the user as a lone achiever in a fair contest. The relational layer (who inside this product would notice if I left, who do I check on, who is counting on me Tuesday) is usually the actual retention engine, and it appears on no dashboard, because relations are not events you can log. We measure the Jake layer. We are retained by the Amy layer. The industry then draws the obvious wrong conclusion from its own instruments: double down on what the dashboard can see.

The uncomfortable half of the diagnosis is that this is the same blindness Gilligan documented in 1982 and Reilly documented in engineering orgs: a measurement system built around legible individual achievement will first ignore care work, then implicitly punish it, then wonder where its most generous contributors went. Your community’s unpaid moderators, your forum’s question-answerers, your guild’s officers are doing the glue work of your retention curve for free, invisibly, at stage two, and stage two always eventually quits. The elephant is not that care is missing from our products. It is that care is already carrying them, unmeasured, unthanked, and undesigned.

How to Apply the Ethics of Care with the Octalysis Framework

Gilligan gives you the moral architecture: two voices, three stages, and a completion condition (care must be received, not just performed). What she does not give you is the motivational engineering: which drives to pull, in what order, to make a care loop people actually want to inhabit. That is the gap the Octalysis Framework fills, and the crosswalk below is, as far as I know, the first systematic mapping between the two.

The Web of Relationships: Core Drive 5 + Core Drive 1

Care ethics relocates the unit of moral analysis from the individual to the relationship, and the design translation is direct: stop asking “what does my user want?” and start asking “what relations does my product let my user maintain?” The engine here is Core Drive 5 (CD5): Social Influence & Relatedness, but notice that most features filed under CD5 are actually comparison mechanics: leaderboards, follower counts, status tiers. Those are the justice voice wearing a social costume. Ranking is about separation. Care-frame CD5 is about attachment: mentorship pairings, check-in rituals, the guild slot only you fill, the newcomer whose first week you shepherd. Layer Core Drive 1 (CD1): Epic Meaning & Calling on top and the relation gains a vertical dimension: I am not just connected, I am a caretaker of something beyond myself, whether that is a patient, a protégé, or a commons like Wikipedia. CD1 + CD5 is the signature of every community whose members describe leaving as “letting people down” rather than “canceling a subscription.”

The Responsibility: Core Drive 8 + Core Drive 4

Why does being needed retain harder than being rewarded? Because dependency arms the most powerful drive in the framework: Core Drive 8 (CD8): Loss & Avoidance. When someone or something depends on you, leaving stops being a neutral choice and becomes an abandonment, and humans will move mountains to avoid being the person who abandoned. Pair it with Core Drive 4 (CD4): Ownership & Possession (this is my patient, my student, my village, my streak of showing up) and you have the psychological core of every care loop ever shipped. Now, the warning, and it is the sharpest White Hat versus Black Hat line in this entire library: the identical mechanic splits into honest and dishonest versions depending on whether the dependency is real. A hospice volunteer’s commitment and a Tamagotchi’s death-guilt run on the same CD8 circuitry. When Duolingo’s owl mourns your broken streak, or a farming game’s crops wither on schedule, the design is manufacturing a dependent so it can harvest your guilt. The White Hat version reflects care that flows both directions; the Black Hat version simulates a needy other whose need is a growth metric. Players feel the difference eventually, and when they do, the resentment compounds as quietly as the retention did.

The Three Stages as a Player Journey: Core Drive 2 + Core Drive 3

Gilligan’s stages are a progression, which means you can design them as one. Map them onto the player journey: a newcomer is at survival (let them be cared for: onboarding buddies, low-stakes asks, visible safety), a regular crosses into responsibility (give them someone to care for: a mentee, a moderation shift, a corner of the commons), and a veteran graduates into mature care (let them shape the care system itself, and let them visibly rest without penalty). Core Drive 2 (CD2): Development & Accomplishment makes the progression legible: care contributions need progress markers exactly the way kill counts have them, or stage-two invisibility sets in. Core Drive 3 (CD3): Empowerment of Creativity & Feedback makes the care non-scripted: a carer choosing how to help is engaged; a carer executing a checklist is an unpaid employee. And close every loop with Tronto’s fourth phase, the receipt: the mentee’s “I passed because of you,” the patient’s update, the thanked moderator. Care that is never received curdles, in Noddings’ terms, into mere performance, and performed care churns.

The Stage-Two Trap in Design

One more crosswalk insight, because it predicts a failure you have seen. A care system with rewards pointed only outward (recognition for sacrifice, badges for hours given, praise for always being available) is a stage-two machine: it selects for self-erasing contributors and runs them to exhaustion. The fix is the same one Gilligan found in her interviews and Klimecki found in the scanner: build the carer’s replenishment into the loop. Rest that does not break the streak. Care budgets that are visibly finite and honored. Recognition for sustainable contribution rather than maximal contribution. Communities that survive a decade do this on purpose or stumble into it; communities that implode usually optimized their most caring members like a consumable resource.

The Warning Light on Your Dashboard

Every framework in this library earns a diagnostic, a metric pattern that tells you the principle is being violated before the damage shows up in revenue. Here is Gilligan’s. Watch for individual engagement metrics climbing while relational metrics stay flat: sessions and streaks up, but mutual interactions, repeat collaborations, newcomer survival, and contributor tenure unmoved. That divergence means you are filling the achievement layer while the care layer empties, and the care layer is where long-horizon retention lives. The reverse signature is just as telling: communities where measurable engagement looks mediocre but members keep showing up for years almost always have a dense, invisible web of mutual responsibility your instruments cannot see. Two more lights worth wiring up: a rising share of care work concentrated in a shrinking set of users (your stage-two volunteers are approaching burnout, and they will not warn you before they go), and “guilt-flavored” return sessions, users who come back to service a dependency mechanic and immediately leave. The first is a staffing crisis in costume. The second means your CD8 loop is harvesting compliance, and compliance churns the moment a competitor offers the same value without the leash.

Practical Steps for Designing with Care

Here is the working checklist I would hand a product team Monday morning:

1. Audit your social features with the two-voice test. For each one, ask: does this connect users by comparison (ranking, status, fairness) or by attachment (dependency, response, mutual aid)? If your entire social layer is comparison, you have built Jake’s product and Amy’s users have nothing to stay for.
2. Give your user someone or something to care for. A mentee, a team slot, a shared resource, a living thing, a corner of the commons. Make the dependency real: something that does measurably better because they showed up.
3. Close the Tronto loop. Every care contribution needs a path for the response to come back: the thank-you, the progress update, the visible effect. Unreceived care is the most common silent killer of contribution systems.
4. Make care work legible. Whatever your platform’s equivalent of glue work is (answering newcomers, moderating, maintaining), give it the progress markers and recognition surface that individual achievement already enjoys. Measure the Amy layer.
5. Design for stage three, not stage two. Build rest, limits, and replenishment into the carer’s loop. If your recognition system rewards always-available self-sacrifice, you are strip-mining your best people.
6. Run the White Hat check on every dependency mechanic. Is the need real, or manufactured so its abandonment can be punished? If the cared-for entity exists primarily to generate guilt, you are renting retention at the price of trust.
7. Interview for the question behind the answer. In user research, code the moral architecture, not just the choice. “Why did you stay?” answered with names and obligations is a different signal than the same retention event answered with features and prices.

The Different Voice Was the Beginning, Not the End

Gilligan’s theory took real damage and earned real vindication, usually in the same decade. The gender claim shrank to a small effect size. The care voice itself, though, escaped the gender war and conquered territory she never aimed at: nursing codified it, philosophy systematized it, neuroscience found its circuitry and even its burnout signature, and organizational research rediscovered her stage-two trap inside every promotion rubric that cannot see glue work.

For behavioral designers, I would compress her legacy into one sentence: people are retained by what they care for, not only by what they achieve, and a design that cannot hear the care voice is deaf to its own strongest force. The justice voice gave us points, ranks, fairness, and progress, and we built an industry on it. The care voice is still mostly unmined. Somewhere in your product right now there is an Amy, holding a web of relationships together, invisible to your dashboard. The next decade of behavioral design belongs to whoever builds for her on purpose.

Frequently Asked Questions

What is Gilligan’s ethics of care in simple terms?

The ethics of care is Carol Gilligan’s theory that moral problems are about preserving relationships and responding to needs, not only about applying fair rules. She argued psychology had treated justice-style reasoning as the gold standard and wrongly scored care-style reasoning as immature.

What are Gilligan’s three stages of care?

Stage one is orientation to individual survival, where care aims at the self. Stage two is goodness as self-sacrifice, where care aims at everyone except the self. Stage three is the mature ethics of care, where self and others are held in one web under a principle of hurting no one, including yourself.

What is the difference between care ethics and justice ethics?

Justice ethics reasons from separation: individuals with rights, conflicts settled by ranking principles and applying rules impartially. Care ethics reasons from connection: people embedded in relationships, conflicts settled by tracing needs and responding without abandoning anyone. Mature moral agents use both.

Who were Jake and Amy?

Two eleven-year-olds in Gilligan’s research who answered the Heinz dilemma differently. Jake treated it as a logic problem (life outranks property, so steal). Amy traced the relationships (prison leaves the wife uncared for, so talk to the druggist). Kohlberg’s manual scored Jake a stage higher; Gilligan argued Amy was answering a different, equally mature question.

Did research confirm that women and men reason differently about morality?

Mostly no. Meta-analyses, including Jaffee and Hyde’s 2000 review of 113 studies, found only small gender differences in care and justice orientation, with the dilemma’s content predicting the moral voice better than gender does. The care orientation is real and undervalued, but it is human rather than female.

Is the ethics of care a feminist theory?

It emerged from feminist psychology, but feminists are also its sharpest critics. Scholars like Catharine MacKinnon warned that celebrating care as a women’s voice risks dignifying the unpaid caring roles women were confined to. Later care ethicists like Joan Tronto reframed care as a political, fully human practice rather than a gendered trait.

How does Gilligan’s theory relate to Kohlberg’s stages?

Gilligan trained under Kohlberg and built her theory as a correction to his. His stages track the development of justice reasoning; hers track the development of care reasoning. The modern consensus treats them as parallel moral tracks rather than rivals, each with its own mature and immature forms.

What did Nel Noddings add to the ethics of care?

Noddings supplied the philosophical structure: the unit of care is the relation, not the carer’s virtue, and a caring relation is complete only when the cared-for receives and responds to the care. Unreceived care is incomplete, which has direct consequences for designing contribution and community systems.

How do designers apply the ethics of care?

Audit social features for attachment versus comparison, give users someone or something real to care for, close the feedback loop so care is visibly received, make care work as legible as individual achievement, and build rest into the carer’s loop so the system selects for sustainable care rather than self-sacrifice.

Which Octalysis Core Drives map to the ethics of care?

The relational web maps to Core Drive 5 (Social Influence & Relatedness) elevated by Core Drive 1 (Epic Meaning & Calling). Responsibility and dependency run on Core Drive 8 (Loss & Avoidance) plus Core Drive 4 (Ownership & Possession). The three-stage care progression is scaffolded by Core Drive 2 (Development & Accomplishment) and Core Drive 3 (Empowerment of Creativity & Feedback).

References

• Gilligan, C. (1977). In a different voice: Women’s conceptions of self and of morality. Harvard Educational Review, 47(4), 481-517.
• Gilligan, C. (1982). In a Different Voice: Psychological Theory and Women’s Development. Harvard University Press.
• Brown, L. M., & Gilligan, C. (1992). Meeting at the Crossroads: Women’s Psychology and Girls’ Development. Harvard University Press.
• Kohlberg, L. (1981). Essays on Moral Development, Vol. 1: The Philosophy of Moral Development. Harper & Row.
• Kohlberg, L., & Kramer, R. (1969). Continuities and discontinuities in childhood and adult moral development. Human Development, 12(2), 93-120.
• Noddings, N. (1984). Caring: A Feminine Approach to Ethics and Moral Education. University of California Press.
• Tronto, J. (1993). Moral Boundaries: A Political Argument for an Ethic of Care. Routledge.
• Held, V. (2006). The Ethics of Care: Personal, Political, and Global. Oxford University Press.
• Walker, L. J. (1984). Sex differences in the development of moral reasoning: A critical review. Child Development, 55(3), 677-691.
• Thoma, S. J. (1986). Estimating gender differences in the comprehension and preference of moral issues. Developmental Review, 6(2), 165-180.
• Jaffee, S., & Hyde, J. S. (2000). Gender differences in moral orientation: A meta-analysis. Psychological Bulletin, 126(5), 703-726.
• Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An fMRI investigation of emotional engagement in moral judgment. Science, 293(5537), 2105-2108.
• Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R. J., & Frith, C. D. (2004). Empathy for pain involves the affective but not sensory components of pain. Science, 303(5661), 1157-1162.
• Decety, J., & Jackson, P. L. (2004). The functional architecture of human empathy. Behavioral and Cognitive Neuroscience Reviews, 3(2), 71-100.
• Klimecki, O. M., Leiberg, S., Ricard, M., & Singer, T. (2014). Differential pattern of functional brain plasticity after compassion and empathy training. Social Cognitive and Affective Neuroscience, 9(6), 873-879.
• Eisenberg, N., & Lennon, R. (1983). Sex differences in empathy and related capacities. Psychological Bulletin, 94(1), 100-131.
• Haidt, J. (2012). The Righteous Mind: Why Good People Are Divided by Politics and Religion. Pantheon.
• Chou, Y. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media.

Related Reading

• Kohlberg’s Stages of Moral Development: the justice-voice counterpart to this guide
• Haidt’s Moral Foundations Theory: where Care/Harm sits among the five foundations
• The Octalysis Framework: the complete guide to the 8 Core Drives
• The Behavioral Framework Library: every framework pillar in one place

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?